Instability on launch with asparagus staging

[Starchaser]

I've recently begun experiencing instability in my rocket launches. I'm 150 days into a career, and I often have 5+ active flights at the same time, so it's not that I'm doing anything wrong on my launches. Within the last few days I've been trying to launch and if I fail, I've been letting Mechjeb's autopilot do it. (which it doesn't seem is working properly. It won't throttle down to avoid going over terminal velocity.) That's been hit or miss too, but I did have it recover a launch where the rocket did a loop.

This morning, I was launching a new station for a mission, and instead of docking in orbit I decided to launch it at 1 shot.To make a long story short,  I launched it with 6 boosters in asparagus staging. Everything was fine during the initial stage, but went crazy after dropping the first pair of boosters. It did a loop and then started spinning like clock hands. It reached a peak height of about 26km. It dropped the second pair of boosters at around 12km, and remarkably they didn't crash into anything. By 8km, between rocket gimbal and RCS thrusters, mechjeb managed to get the rocket stable and pointed in the right direction, and successfully put the station into orbit, with barely enough delta-v to make the Mun transition and then do an insertion burn. (at least, I assume it will be enough. It's still in transit, but I have 750 m/s left

That event convinces me that I'm doing something wrong in setting up my asparagus staging. I'm hoping someone can help me with what I'm doing wrong.

Here come the pictures

Long shot of rocket. I actually can't pull back far enough to get the entire rocket in the shot:

https://i.gyazo.com/37d143a688df596f9159685bf21c6ec8.jpg

Close up of boosters

https://gyazo.com/20112520147c87a577a023c8c3d68aca

Top down: 

https://gyazo.com/a0e8b00af0446c6ec32d756d12462cec

Angled a bit to see stitching:

https://gyazo.com/588592116f365aedf5365c93bbf030b9

Thanks in advance for any assistance

[Spricigo]

Seems you already solve the problem. But in a 'brute force' approach, so I suppose you may still be interested in what exactly happened and a reliable way to avoiding it infuture. SO:

CoM not in centerline (look at torque values given by KER)

Almost certainly the culprit is the payload, try to make it as symmetrical as possible. Don't be ashamed to use the offset tool. Also, if possible, use fairing to avoid asymmetric aeroforces.

[Superfluous J]

That rocket displays almost none of the typical problems, other than seeming pretty beefy for its payload. The TWR isn't *too* bad really, but it's a shade on the high side for my tastes.

Also, terminal velocity really doesn't matter since 1.0 came out. Used to be going over it was super inefficient. These days, if you're not going too fast until you explode. Mostly.

Anyway, the first thing I think of is maybe your boosters are stable because they have this nice ring of struts holding them together, and that ring is snapped when you drop the first boosters, revealing an instability. Maybe strut to the center stack instead, or use autostruts to grandparent?

I won't be able to do it personally (busy weekend), but someone else may be able to load your craft file and see something more by actually flying it.

[Starchaser]

17 minutes ago, 5thHorseman said:

Anyway, the first thing I think of is maybe your boosters are stable because they have this nice ring of struts holding them together, and that ring is snapped when you drop the first boosters, revealing an instability. Maybe strut to the center stack instead, or use autostruts to grandparent?

Made me go back in and look. I could have sworn the boosters were strutted to the center stack, and they are, both at the top and the bottom. However, it's kind of an oblique strut instead of more direct. Maybe I moved the part after settng the first strut.Unless I strutted it in the opposite direction too, it might act like it's unstitched from force in a certain direction. I'll redo that stitching and try the launch again

 

Edit: I removed 4 of the boosters to look at the original set It was stitched at top and bottom and left and right, to the center stack.

https://gyazo.com/02672a0b4a2c4c295e6b409e4ecf0845.

I'm going to try to shorten them some and try to launch again

 

Edited October 22, 2017 by Starchaser
Updated info

[Starchaser]

21 minutes ago, Starchaser said:

I'm going to try to shorten them some and try to launch again

New stitching,

https://gyazo.com/5197f03ddcf9ed407c66672f7234b603

 

[Starchaser]

Nope.

Mechjeb and I both took 2 or more attempts to launch, because separating boosters took out parts of the rocket. (which seems to be a recurring issue. using the manifold with the 450 ejection force or the high profile decoupler I use here.) Mechjeb wouldn't gain stability once it lost it, until I dropped the final 2 boosters manually.With just the main stack it didn't have enough delta-v to get to orbit. When I piloted it, once I dropped the first boosters, as soon as I touched the control it went crazy.

I finally succeeded in getting it to orbit bugs bunny style. I went straight up and once I dropped the first and second booster sets, I made a left turn at albequerque. I wound up in a circular (ish) orbit at 138km, and 2.5° of inclination, using 4k m/s of delta-v (I could have done a bit better. I was so intent on watching time to apoapsis to keep it from climbing to 400k or some ridiculous number, I missed when the final 2 boosters ran out of fuel, for a bit. The main tank was about 80% full when I did separate.

So, the stitching didn't help my stability. Neither did Jeb trying to sneak into my command pod!

[Kryxal]

If you have trouble peeling off sets of boosters, just use the separatrons, it's what they're there for.  Just how do they crash into it?

[Starchaser]

37 minutes ago, Spricigo said:

Almost certainly the culprit is the payload, try to make it as symmetrical as possible. Don't be ashamed to use the offset tool. Also, if possible, use fairing to avoid asymmetric aeroforces

Possibly, except for 2 things. 

1. I've launched worse without an issue.

2. the torque values go up as you peel off boosters, at it's worst when I'm just having the main stack. It's also at this point that the autopilot regains control.

However, my saying no is just theoretical. I'll pull off the station and replace it with a module and equipment of approximately the same weight, and 0 torque and see if the issue fixes itself. Although I'm not sure where the asymmetry is coming from. All parts were added with the symmetry tool, except for the KER amd mechjeb items, which don't affect torque. My in-line docking port maybe?

Update 1. The 'asymmetry' was mostly in the second PPD-24 Itinerant module.  I have asymmetry in quotes because I pealed off every component on that module and then put it back on, and the 'culprit' was the HG-5 high gain antennas I took it off and the torque on the payload went to 0. I put it back on and the torque stayed at 0. There is some asymmetry, oddly enough on the central stack. Maybe some stitching is off. At any rate, as you can see, the torque values are all less than 0.1 across the rocket now.

 

https://gyazo.com/587ee6dbfa59802e17007bef6d28535d

So let's see how it flies now.

[Starchaser]

Mia culpa. The ascent autp[ilot still couldn't handle it, but it behaved very nice for me. as long as I only tipped less than 10° before dropping the first booster set. After that it did the gravity turn very well. I'll have to pay more attention to the torque values, I normally don't look at them unless I know I'm installing something that is asymmetric. (My rover rocket, for example, has the small I-beam running along it's backside to offset the rover)

2 hours ago, Kryxal said:

If you have trouble peeling off sets of boosters, just use the separatrons, it's what they're there for.  Just how do they crash into it?

Not enough ejection force. The high profile ones only push with 260. The hydraulic manifold is low profile so everything is too crowded close to the stack and things go bump because there's no room. If I'm going nearly straight up they'll fall away mostly ok, but tilted, gravity pulls the number 6 booster into the number 1 booster and knocks it off the decoupler, or it smashes into the main stage, and I discover that when I get to just the main stage, there's no engine on the bottom of the main stack anymore.

I never understood how the separatrons worked until I watched a non-Manley asparagus staging video last night. I'll add them next time. i didn't want to tear this existing rocket down and add them, thinking if I reconnected the second and third booster sets and inadvertently fixed my issue, I wouldn't be able to track it down. Now that I have some idea what was wrong, I'll look at separatrons. 

[Harry Rhodan]

You also keep your CoM lower than it needs to be by attaching the boosters at the very bottom of you center stack.

Edited October 22, 2017 by Harry Rhodan

[Starchaser]

and effects of me waking up at 4 am. those were supposed to be separate posts.

Just now, Harry Rhodan said:

You also keep your CoM much lower than it needs to be by attaching the boosters at the very bottom of you center stack

That was to prevent my "boosters crashing into rocket" issue as much as I could. I'll raise them again if I can figure out how to make the separatrons work for me.

[Palaceviking]

Quite a fat rocket with a far back centre of mass, it's aerodynamics and balancing will be a bit much for mechjeb,  I would try to get more stuff for in-line,  I know asparagus will save you a little Dv but instability is one reason it's not used irl. 

[FancyMouse]

7 hours ago, Starchaser said:

Not enough ejection force. The high profile ones only push with 260. The hydraulic manifold is low profile so everything is too crowded close to the stack and things go bump because there's no room.

I've never trusted hydraulic manifold exactly because of that. When I separate any stack >= 2.5m I directly go with TT-60 + sepratrons without hesitation. The extra room not only allows a clear separation, but also allows room for 6x symmetry when side stacks are of the same size as central one (the distance between two adjacent side stacks will be exactly the width of the decoupler, and the closer they are, the more likely something bad will happen)